Interview with Senior Director of 4C Supply®, Michael Carnes

As modern clinical trials get more complex and the pressure to prevent waste and stock-outs continues to mount, sites require a system that is simple, flexible, and insightful. 

As Senior Director of 4C Supply®, Michael Carnes has led the development of the clinical supply optimization tool to address the obstacles being faced by today’s clinical supply managers, his insights into the software and its advancement are provided in the following Q&A.

What has your experience been like since you started leading 4C Supply®?

When our CEO, Dave Kelleher, offered me the opportunity to lead 4C Supply®, I was excited at the chance to lead the team and drive overall product development. I had worked in technology roles for over 30 years, so I was in a prime position to focus on building a sustainable, agile software operation to continue the strong foundational work that was set by the product launch team at 4G Clinical. With two years in the position, I’ve learned a lot and the forecasting tool has matured significantly in that time, so I am excited to see the tangible, value-add enhancements that are right around the corner.

Can you give insight into the initial launch of 4C Supply® and how we utilized points of differentiation from 4G Clinical’s experience in RTSM?

Early releases of 4C Supply® were targeted at pre-study forecasting and built to leverage our expertise in Natural Language Processing (NLP). 4G Clinical’s Prancer RTSM® and 4C Supply® use NLP technology to read and interpret written specifications to build a deployable system in moments–with the click of a button. 

In comparison, other processes for building these systems rely on clinical study teams approving these same specifications—hundreds of pages of complex technical requirements—that they may not understand. This process can ultimately lead to unexpected system configurations at UAT, which can cause further delays and costs. With NLP, the system build process allows for an agile approach, provides the study teams with early system access, supports the ability to pivot, ensures the desired end result, and not to mention… all in a fraction of the time. 

How has 4C Supply® evolved since the early versions of the product?

Though we retain an easy-to-learn, easy-to-build NLP framework, we have transitioned the tool to support the needs of large pharma with more sophisticated risk-management processes. We have partners now who want to put 50-100 studies per year in the system. The bar has been raised in terms of study complexity that needs to be built into modeling, performance and accuracy requirements, and the ability to fit into complex supply chain processes. 

Implementing a variable forecasting algorithm was a key building block in this transition. Many forecasting solutions still rely on simulation to track variability. Such solutions can be resource intensive and expensive; users often reduce the number of simulations to make them performant. Leveraging straight-forward distribution mathematics, 4C Supply® offers the equivalent of infinite simulations so supply managers can fully assess risk in their models and assumptions. 

With the combination of a variable forecasting algorithm, powerful modeling capabilities, and actuals integration, we’ve opened the door to automating key aspects of supply management. With 4C Supply®’s next two releases, you’ll be able to read an actuals file each day, get a notification if the forecast varies substantially from the actual results, and automatically re-weight assumptions between your actuals and forecast. The system is evolving to the point where it can support the most advanced forecasting cycle management. 

What does the future hold for 4C Supply®?

Despite the added sophistication and capability, 4C Supply® will stay true to its roots of putting control in the hands of our users. Most supply groups are quite small when compared with clinical operations teams. Providing an elegant solution that lets supply managers easily update their model and assumptions, while also having powerful enterprise-scale features, underpins my product decisions. I don’t want a solution that requires a team of high-priced consultants to create, modify, and run a forecast. However, I want the tool to be smarter, to do things automatically, and to ultimately help supply managers with decision support. Study designs and supply chains are getting more complicated. The trick is to address that complexity but remain easy for an end user who may only use the system on a monthly or quarterly basis. To quote Steve Jobs, “Simple can be harder than complex.”